Risk Analysis of Transport Requalification Projects in the Urban Mobility Problem Caused by a Mining Disaster

Abstract

Background: This paper proposes a risk analysis of transport requalification projects in the urban mobility problem caused by a mining disaster related to irregular rock salt extraction in the city of Maceió, Brazil. Methods: The model is composed of three main steps: problem definition, risk management, and decision analysis. For this purpose, we used the Picture Fuzzy-Delphi method for data collection and experts’ judgment elicitation and the Delphi method was used to assess the problem without interference from others. In addition, we used Picture Fuzzy Sets (PFSs) to incorporate uncertain information in the decision-making process. Results: The results of the proposed model demonstrated consistency and relevance to the discussion. The application of methods shows the risks of the project based on a general perspective. It evaluates the sustainability tripod: economic, environmental, and social points of view, assessing the occurrence risk and intensity of the risk. Conclusions: The main objective of the work was achieved; however, some limitations of this study are related to the methods used to assess risks and the options of projects of requalification available at the moment of data analysis. This paper contributes because it systematizes the risk management of projects related to requalification in urban mobility.

1. Introduction

Cities are complex systems relying on critical services delivered by multiple physical infrastructure networks [1]. Mobility represents the ability of a transportation system to provide ease of movement from one location to another [2]. There is a connection between transport system components and satisfaction related to the quality of life, which is highlighted as follows: (a) the physical layout of the transportation system; (b) the mobility of people from one place to another; and (c) safety in terms of physical conditions and human behaviour [3]. Thus, urban mobility systems are important to society, changing transit times and urban connectivity [4,5]. However, cities in developing countries have a less reliable provision of transport services [1]. This lack of mobility is aggravated in cases of extreme events or disasters.

A disaster is an unexpected event that causes destruction to the affected area, interrupting the functioning of a community and causing material, economic, and human damage, a fact which prevents autonomous recovery [6,7]. A lack of mobility is likely to reduce people’s liveability in external zones, mainly in terms of weakening community networks [8].

In recent decades, there has been an increase in man-made disasters due to oversights from organisations that have had a significant humanitarian impact. Events such as the terrorist attacks of September 11th in 2001 and Hurricane Katrina in 2005 in the United States, the 2004 tsunami in Asia, the 2009 H1N1 world pandemic, the earthquakes in Chile and Haiti in 2010, the Ebola virus in West Africa between 2013 and 2016 and, more recently, the new coronavirus pandemic, beginning in 2019, and the Russian–Ukrainian war, in 2022, are all examples of disasters with long post-disaster environmental, social and economic recovery processes [9,10,11,12].

More specifically, in Brazil, the disaster caused by irregular rock salt extraction destabilised the subsoil and sank five neighbourhoods in the capital of the state of Alagoas, the city of Maceió, Brazil [13]. Since 2018, the city has faced significant challenges in terms of social, physical, and economic requalification. Considered one of the most significant ongoing disasters occurring in an urban area, among the main impacts is the problem concerning urban mobility, primarily due to the stoppage of part of the light rail vehicle line, which consequently also caused damage to the mobility of both people and vehicles due to the closure of essential roads in the city of Maceió. Despite there being alternatives to conventional transportation services, such as cars and bike-sharing programs, considering sustainable concerns, it is necessary to create a structure that enables these new transport models to operate safely. This mobility structure involves more bike lanes and alternative roads to divert heavy vehicle mobility caused by the interruption to the roads in the affected neighbourhoods. For [14], mobility is considered essential in enabling well-being and connecting people, places, and possibilities.

Thus, urban resilience has gained significant attention as a new risk management and disaster mitigation paradigm [15]. Urban resilience refers to the ability of an urban system to maintain or rapidly return to desired functions in the face of a disturbance, to adapt to change, and to quickly transform systems that limit current or future adaptive capacity [16]. Many authors have studied the urban mobility problem post-disaster to reduce mobility congestion and quickly respond to the needs of the affected population [1,15,17,18]. On the other hand, our work assesses risk management in terms of requalification projects for transport systems affected by a disaster that made vehicle mobility worse throughout the city.

Therefore, this paper proposed a risk analysis of transport requalification projects related to the urban mobility problem caused by a mining disaster related to rock salt extraction in the city of Maceió, Brazil. The model consists of three main steps: problem definition, risk management, and decision analysis. The Picture Fuzzy-Delphi method was used for data collection and experts’ judgment elicitation. The Delphi method allows experts to assess the problem without interference from others [19]. Picture Fuzzy Sets (PFSs) can incorporate uncertain information and make a decision-making process more realistic [5]; furthermore, a descriptive statistic for the concordance and discordance analysis among the specialists and a Copeland matrix to obtain a rank position and a score for each requalification project, were used. Copeland is adequate in terms of PFSs when we are challenged with human opinions involving a greater variety of answers: yes, abstain, no, or refusal [20]. The real-life case of the urban mobility problem in the municipality of Maceio, Alagoas, was studied. For this purpose, four options for requalification projects were evaluated by seven experts.

Post-disaster requalification of transportation infrastructure is a complicated task because it involves high investments, changes in the present mobility options, and the creation or improvement of infrastructure [21]. In general, the local government does not have the availability of financial resources to provide such improvements. Therefore, models, such as the one proposed in this paper, can contribute to better decision-making procedures, supporting decision-makers. Thus, the main contribution of this paper is to provide a systematised model, with well-defined steps, for decision-making related to the requalification of urban mobility considering the necessary post-disaster actions. It can be helpful for policymakers and public managers and enable them to make more consistent decisions.

In addition to this introduction, the paper is organised as follows. Section 2 introduces a theoretical background. Section 3 presents the proposed model. A case study is presented and discussed in Section 4. Finally, in Section 5, the conclusions and implications of the research are assessed.

2. Theoretical Background

2.1. Risk Management

Disasters place extraordinary demands on an affected region’s corporate and governmental organisational capacities [22]. Disaster management covers three main phases: disaster preparedness, disaster relief, and disaster recovery and requalification.

Disaster preparedness focuses on community resilience and disaster prevention capacity in disaster-prone areas [23], enabling people to reside safely and peacefully in their environment [24]. The social and financial impacts of extreme events demand an analysis of their causes in order to improve mitigation strategies and the preparedness of individuals and communities to reduce disaster effects [25]. In this context, authors have studied interventions designed to reduce disaster risk [23,24,26,27,28,29,30,31].

Disaster relief, i.e., humanitarian operational actions concerned with the immediate needs of victims, recognises logistics as a vital feature of disaster relief operations [27,32,33]. The authors mention humanitarian logistics as a crucial strategy in reducing the negative impacts on the population affected by the disaster. Humanitarian logistics manages several activities related to requalifying transportation mobility and infrastructure, providing supplies, and organising assistance to the affected people. Thus, most studies are related to identifying and analysing emergency logistics risks, where transportation is a crucial operational element [34,35,36,37].

Disaster recovery and requalification are long-term actions involving the planning and managing strategies for situations aimed at community rehabilitation (post-disaster). For this, the requalification of transport infrastructure is crucial, and a high priority must be given to avoiding substantial economic losses by restoring damaged transportation infrastructure [21,38]. Therefore, the study focuses on reducing the risk of potential disasters [21,39,40].

However, there are gaps regarding urban transport system requalification in post-disaster strategies caused by human actions or omission. The relevance of specific studies comes from the need to analyse the risks associated with transport restructuring alternatives (projects).

2.2. Fuzzy Approach

Fuzzy sets are widely used in the literature, and urban mobility problems (UMPs) are no different. We can report some studies, such as [5], which developed an integrated model based on the Fuzzy analytical hierarchical process (FAHP) in relation to urban mobility system (UMS) planning. Ref. [41] used Fuzzy Functions in their prioritisation model of urban areas for integrated revitalisation and sustainable urban mobility (SUM) projects. While [42] proposed the Fuzzy-FUCOM and Fuzzy-CoCoSo methods to select measures and policies to be carried out to achieve SUM plans. With a similar goal, Ref. [43] applied the Fuzzy D Dombi (Fuzzy 2D) algorithm to evaluate the alternatives of transition from the traditional approaches to sustainable urban mobility.

Ref. [44] proposed a neuro-Fuzzy inference system (ANFIS) to create urban mobility estimation based on the telecommunication activities within public mobile telecommunication networks. Ref. [45], based on the pressure–state–response (PSR) model with 25 indicators, used the Fuzzy comprehensive evaluation (FCE) method to assess the overall status of urban mobility.

Ref. [46] proposed the Fuzzy Dombi-based Combined Compromise Solution (D’ CoCoSo) to prioritise circular economy concepts for urban mobility. Ref. [47] proposed the Fuzzy Einstein Weighted Aggregated Sum Product Assessment (WASPAS) approach for the economic and societal dynamics of climate change mitigation strategies in urban mobility planning.

However, there are decision contexts in which the decision maker does not know how to evaluate or considers that there is no relationship at some point. There are Picture Fuzzy Sets (PFSs) for these situations.

PFSs are direct extensions of the Fuzzy sets and intuitionistic Fuzzy sets (IFSs) [20]. It considers the concept of an element’s positive, negative, and neutral membership degree [48]. The PFS approach can provide a consistent valuation based on a simple evaluation from the decision parties [49].

A PFS A on a universe x is mathematically represented as Equation (1) [50].

$$A=\left\{\left(x,{\mu }_A\left(x\right),{\eta }_A\left(x\right),{\nu }_A\left(x\right)\right)\right\}$$

(1)

where ${\mu }_A\left(x\right)$ is the degree of positive membership of x in A; ${\eta }_A\left(x\right)$ is the degree of neutral membership of x in A; and ${\nu }_A\left(x\right)$ is the degree of negative membership of x in A.

There is an essential relationship between these terms: ${\mu }_A\left(x\right)+{\eta }_A\left(x\right)+{\nu }_A\left(x\right)\le 1$. The difference ${\pi }_A\left(x\right)=1-{\mu }_A\left(x\right)+{\eta }_A\left(x\right)+{\nu }_A\left(x\right)$ represents the degree of refusal membership of x in A.

An essential point of PFS is defining the preference relationship between Picture Fuzzy numbers. If a and b are two Picture Fuzzy numbers, we can define two measures called Score (S(x)) and Accuracy (A(x)). After that, these relations are validated

$$\mathrm{If} S\left(a\right)>S\left(b\right), \mathrm{then} a>b$$

(2)

$$\mathrm{If} S\left(a\right)=S\left(b\right) \mathrm{and} A\left(a\right)>A\left(b\right), \mathrm{then} a>b$$

(3)

In this case, “>“ means the preference between PFS numbers.

2.3. Copeland Method

The Copeland method is a well-known methodology used in terms of social choice. It is a voting system that offers a conceptual basis for determining the rank or preference of candidates in an election [51]. It is a straightforward and intuitive method that relies on pairwise candidate comparisons. While it does not assign specific scores or weights to candidates, it calculates a measure known as the Copeland score, which represents the number of pairwise victories a candidate has over other candidates [52].

A Copeland Matrix can be defined as $C=\left[c_{ij}\right]$. If $c_{ij}=1$ it means that the alternative $i$ wins $j$. Column W and row L count the number of wins and losses. Copeland’s Score represents the total of wins minus losses; the largest score means a more consistent ranking [53,54]. In

Table 1. Copeland matrix computation.

	P1	P2	P3	P4	W	Copeland Ranking	Copeland Score
P1		1	0	0	1	3	−1
P2	0		0	0	0	4	−3
P3	1	1		1	3	1	3
P4	1	1	0		2	2	1
L	2	3	0	1

Source: The authors (2023).

, we present an example of Copeland computation.

In the Copeland matrix, if the row alternative has priority to the column alternative, the number $1$ is placed in the related element of the matrix. If the row alternative does not have priority over the column alternative, the number $-1$ is placed in the related element. Also, if the row alternative and column alternative are equal, the number $0$ is placed in the element [55]. The sum of the rows represents the final score of the element. Thus, the Copeland method presents a partial ranking.

There is an intuitive connection between PFS concepts and voting systems, as in the Copeland method. Ref. [52] commented that Picture Fuzzy Sets are an efficient tool for dealing with uncertainty and vagueness, particularly in situations requiring the assimilation of more linguistic assessment dimensions such as human voting, feature selection, etc. Fuzzy logic allows for the representation of uncertainty or imprecision in decision making.

In the context of the Copeland method, a PFS can be used to represent the pairwise comparisons among candidates. Instead of crisp victories or defeats, Picture Fuzzy Sets can capture the degree of preference or support between candidates. These characteristics allow for a more nuanced evaluation of the pairwise comparisons, accounting for uncertainty and imprecision. This process helps decision makers in expressing their preferences by voting more naturally [56].

In the literature, some authors used an integrated multicriteria group decision-making methodology (MCGDM) considering the Copeland method, such as in [53]. However, its relationship with PFS is new and primarily related to the problem proposed in this article.

3. Methodology

The proposed model aims to evaluate the project risks and provide relevant information based on a group of specialists, i.e., a group decision-making situation. One of the advantages of this method is a simplified input method to obtain information from the specialist. As illustrated in Figure 1, the proposed model has three main steps: (i) problem definition, (ii) risk management, and (iii) decision analysis.

In Step 1, related to the problem definition, we characterised the urban mobility problem in addition to defining the selection of experts and the project variables. In Step 2, the risk management involved identifying the set of risks and the categorisation in terms of economic, environmental, and social dimensions of sustainability; then, we assessed the risks based on the Picture Fuzzy Delphi method, generating the probability and intensity of risks.

Finally, in Step 3, we conducted the decision analysis, applying the concordance and discordance analysis and the Copeland matrix, which provided scores and rankings of the risks. Therefore, based on that, we analysed the results and provided some recommendations based on these steps, which are well described in the next section.

The proposed method introduces some advantages in terms of risk analysis. Firstly, a structured questionary based on Fuzzy numbers can easily elicit the process’s inputs, as in [49]. This is an advantage over the other methods that require complicated elicitation. The Delphi method based on Fuzzy processes increases the robustness of the evaluation. Lastly, the use of the Copeland evaluation to detect the conflictual evaluation among specialists is a solid method to create a consensual assessment. Models, such as [57], represent a risk ranking, while our proposed method highlights more conflictual evaluations to suggest discussions about the theme to support a robust specialist evaluation. These topics are described in the following section.

3.1. Problem Definition

The objective of this first stage is establishing a correct definition of the problem scenario and the project solution. There are some relevant actions required to define the problem correctly and its parameters.

• Problem scenario: the scenario of the problem and project proposed to handle this situation should be well defined.
• Expert selection: selecting relevant and heterogeneous experts is important to obtain evaluations from different perspectives.
• Project definition: improvement projects are relevant to handling some complex problems. A good description of the problem is a good input for good projects.

The undertaking of projects used to be complex, and the conversion from planning to real-world scenarios has some risks involved. The experts should assess the project risks and the criteria used to evaluate it.

3.2. Risk Management

Risk management procedures have three essential steps: identification, categorisation, and assessment [36]. Identification entails evaluating all potential risks in a given situation. The works of [50,58] help in the UMP field in this step. In addition, specialist opinions are useful at this stage. After that, risks are categorised into homogenous categories. According to the sustainable urban mobility project view, we consider the triple bottom line of sustainability, i.e., social, economic, and environmental dimensions.

Thus, decision makers determine how to respond to categorised risks during risk assessment. This paper used a general procedure based on [49]. We increased a natural connection between PFS and Delphi method, as in the Picture Fuzzy Delphi method (FDM) [59], for the risk assessment stage. “The Delphi survey method has been demonstrated as an effective approach to solicit experts’ opinions on complex issues and achieve long-term projections” [19]. “While the use of fuzzy theory avoids the distortion of individual expert opinions, captures the semantic structure of predicted items and considers the unclear nature of the data collected” [59].

The Delphi method supports the data collection using a self-administered questionnaire composed of closed questions, where the projects (solutions to the problem) were presented briefly. While the Picture Fuzzy Sets (PFS) model allows the decision-maker (DM) to assume that he/she does not know or see any relationship at some point; they cannot evaluate. Thus, the DM defines her/his assessment by assigning one of the impact degrees: high, medium, and low, or a refusal to answer [60,61]. This method limits the risk effects based on two perspectives: risk impact and risk occurrence, as in [62]. Thus, each expert evaluated each projects pi of set P, such that i = {1, 2, …, n}, in each risk $x_j$xj of set A, such that j = {1, 2, …, m}, considering two criteria: ($g_1$₁) Probability of the risk $x_j$ occurrence in a project p_i; and ($g_2$g) Intensity of risk x_j impact in a project $p_i$p.

3.3. Decision Analysis

In the last step, we evaluate the results obtained in the model. First, the model uses descriptive statistics to assess the opinions dispersion. The main objective of this step is to highlight the relations with the most significant level of concordance and discordance among the specialists. The concordance/discordance evaluation is interesting in terms of evaluating the differences between the results and guiding the discussion to reinforce the global group evaluation. It can be performed based on Fuzzy operators combined with statistical concepts.

We propose that the score function $S\left(a\right)$ in the PFS literature is an interesting metric to conserve fair PFS values in the Copeland structure. The score is used to compensate for large values. ${\nu }_A\left(x\right)$ in comparison with ${\mu }_A\left(x\right)$. This process uses more PFS-generated information that could not be considered if the process evaluates just the membership value. Thus, the preference functions (Equations (2) and (3)) are inputs in the Copeland matrix that show the winner in a complete pairwise comparison.

The Copeland method provides a clear conceptual basis for understanding the relative support levels between candidates. It focuses on pairwise comparisons, which are often easier for voters to evaluate than assigning explicit scores. By counting the number of victories, the Copeland method captures a sense of overall preference without relying on complex algorithms or subjective weighting. This method can be defined as follows: for all $R\in B$, as an alternative be determined by the function: $c_{R:}:X\to R_{+}$ such that for all $A\in$ X and all $R^{\prime }\in M_A$, $c_{A:}\left(A\right)=mi{n}_{R^{\prime }\in M_A}\delta \left(R,R^{\prime }\right)$, $A$ is a Copeland ranking if and only if $A’\in A,A{\succcurlyeq }_C{A}^{\prime }\leftrightarrow c_{v\left(u\right)}\left(A\right)\le c_{v\left(u\right)}{A}^{\prime }$ [49]. Thus, we obtain a rank position and a score for each project. The results are analysed and recommendations are made.

One advantage of using Picture Fuzzy Sets in the Copeland method is a more granular analysis of pairwise comparisons. Instead of relying solely on the number of victories, the process can consider the strength or intensity of those victories. This helps distinguish between close and overwhelming victories, providing a richer understanding of the candidates’ positions. However, it is important to note that integrating PSF into the Copeland method may introduce additional complexity. Defining Fuzzy membership functions and handling Fuzzy operations require careful consideration and expertise in Fuzzy logic. Furthermore, the interpretation and manipulation of Fuzzy rankings may pose challenges as they may not always align with traditional crisp rankings.

4. Case Study

4.1. Problem Definition

Here, we are interested in evaluating projects proposed to solve the urban mobility problem in Maceió, Alagoas, Brazil. In March 2018, Maceió registered heavy rains and an earthquake, felt mainly in the Pinheiro neighbourhood. After this event, cracks were observed on streets and buildings. In June 2018, the Geological Service of Brazil, through the Mineral Resources Research Company (CPRM), began investigating the phenomenon and concluded that the problem was the salt extraction mines. Thus, CPRM recommended the evacuation of the area and the creation of a protected area.

In December 2020, the Federal Public Ministry (MPF) and mining company, with the participation of the Public Ministry of the State of Alagoas (MPE), signed an agreement for Socio-environmental Repair. The agreement provides measures to mitigate, repair, and compensate for the impacts of the geological phenomenon in Pinheiro, Mutange, Bebedouro, Bom Parto, and Farol. These established that specific actions for the social-urban recovery in the neighbourhoods are necessary with a focus on improvements in urban mobility and social compensation for this action; the agreement provides for the allocation of the amount of BRL 360 million (as set out in article 63 of the agreement) in addition to the preservation of historical and cultural heritage. It also considered the importance of stabilising deactivated rock salt wells and monitoring geological phenomena and subsidence events in the coming years.

It is worth noting that the municipality does not have an Urban Mobility Plan, as provided for in Federal Law 12,587/2012. Light Rail Vehicles (LRVs) are the only alternative for the people of Maceio to buses, and as a public transport system, offered one of the most affordable rates among Brazilian capitals, had 16 stations and 34 km of lines before the disruption of the stretch.

Among the multiple negative impacts of this man-made disaster, we can highlight the urban mobility problem due to the isolation of the affected area. Thus, 10 bus lines and a section where the LRV passed through were deactivated. The original LRV route, i.e., before the disaster, can be seen in Figure 2 (path in blue). The red circle approximately marks the area affected by the mining disaster.

In this paper, we consider four projects related to Maceio’s urban mobility. First, as a palliative alternative (here named Project 1—P1), the LRV section between Bebedouro and Bom Parto stations, which passed through the area affected by the disaster, was deactivated. Thus, a bus route (in yellow in Figure 3) was intensified to integrate these two stations. Here, also the red circle approximately marks the area affected by the mining disaster.

Despite P1 solving the problem of moving people in the region, it caused a high rate of vehicle congestion, especially on Avenida Fernandes Lima, an important avenue in the municipality, i.e., the problem of urban mobility in Maceió was not resolved. Thus, as a definitive answer, a second Project (P2) was presented. This suggested establishing the same route as used in Project 1 (Figure 3) but making adaptions for the use of a BRT (Bus Rapid Transit) system. However, there are impasses since some roads are not wide enough for the necessary adaptations to use a BRT system; this alternative would probably partially solve the problem.

As a third Project option (P3), we proposed maintaining the original route of the LRVs and raising the rails in the areas where the subsidence of the ground occurred. According to the intersectoral agreement, this area can be trafficked again when the stability of the soil is confirmed. However, according to the Brazilian Urban Transport Company, in Portuguese Companhia Brasileira de Transporte Urbano (CBTU), at this time, it is not possible to guarantee the safety of this area, and there is not even a specific date for the stability of the soil to be confirmed. A follow-up period of 10 years was defined as the time needed for the total stabilisation of the area (as set out in Article 18 of the agreement), this time computed from the closure of all wells, which is scheduled for 2024.

The urban mobility problem in Maceió is urgent and should not wait for the confirmation of soil stability to be resolved. Therefore, a fourth Project (P4) was proposed. This establishes that the route between Bebedouro and Bom Parto stations be carried out again using LRVs with a new route (in yellow), by passing the area affected by the disaster (red circle), as shown in Figure 4. However, this project involves a high investment cost.

The proposed solutions present a series of impasses between the mining company, which must solve the urban mobility problem (UMP), and the government (representative of social interests). Understanding the risks associated with each project can be an essential way to understand the problem and support the negotiation between the stakeholders. It is worth mentioning that all the proposals presented here came from an interview with CBTU members.

4.2. Risk Management

As mentioned earlier, the risk management procedure has three important steps: identification, categorisation, and assessment. In the first step, a set of risks were previously defined based on works such as [50,58]. These were taken to a specialist, a university professor in the field of urban mobility. The specialist was asked to evaluate each risk, responding that he/she considers it a risk when analysing projects for the generic UMP, i.e., the UMP that occurred in Maceió city was not reported to the specialist. This specialist was allowed to insert or exclude information. Then, the risks were categorised into the triple bottom line. As a result, we provide

Table 2. Risks defined for the urban mobility problem.

Dimension	Risks	Description
Economic	Investment	Risk of the project not meeting the budget for its completion
	Maintenance	Risk of the project has a higher maintenance cost than estimated in the project
	Revenue	Risk of the project having a collection lower than estimated by the project
	Price	Completed project risk increasing distance, leading to higher costs to users
Social	Demand	Risk of the project not meeting the needs of the population after completion
	Comfort	Risk of the project not being welcoming and comfortable once completed
	Historical issues	The risk of the project not contemplating initiatives that protect and respect the architecture, history and symbolic value of the place
	Security	Risk of the project increasing user insecurity after its implementation
Environmental	Infrastructure	Risk of the project not guaranteeing the necessary urban infrastructure and urban equipment
	Mobility	Project risk does not improve the flow of public transport once completed.
	Legislation	The risk of the project not complying with the environmental conditions of the project and current legislation.

Source: the authors (2023).

.

4.3. Decision Making

A total of seven experts assessed the risks. The experts were selected for convenience, according to their expertise: university professors in civil engineering or architecture, community leaders in the affected region, and the municipality’s urban mobility department employees. All experts must be residents of Maceio city, so they have some contact with the problem to better understand the projects (alternatives). The analyst followed the Delphi method adapted to the form of the Picture Fuzzy Delphi (PFD) method to clarify the question for the specialists and support them in answering a questionnaire. These processes were undertaken between March and April of 2023. The experts should evaluate all risks according to occurrence and intensity criteria. They should determine if the risk is high, medium, or low or refuse to answer this topic, as mentioned in 3.2.

After the experts evaluated each risk for each project, observing the two defined criteria, the study evaluates the concordance/discordance among the specialist evaluations. To do so, we used a dispersion metric and considered all risks/projects/criterion. Each evaluation was ranked from the more concordant assessment, where the specialist agreed, to those with the largest disagreement. Therefore, for example, if all seven specialists decided that the occurrence of mobility risk in Project 1 is high, this evaluation should receive a ranking of 1. On the other hand, historic issues risk occurrence in P4 has two specialists evaluating it as having high risk, two as medium, two as low, and one refused to answer. The evaluation is wildly divergent. Therefore, the ranking of this evaluation should be high. These results are presented in

Table 3. Concordance and discordance analysis.

		Projects
Criteria	Risk	P1	P2	P3	P4
Occurrence	Investment	4	43	23	66
	Maintenance	4	23	43	66
	Revenue	11	53	77	43
	Price	53	77	38	43
	Demand	18	23	11	77
	Comfort	1	43	18	11
	Historical issues	29	66	53	77
	Security	53	77	4	1
	Infrastructure	66	53	11	38
	Mobility	1	8	18	66
	Legislation	8	38	53	29
Intensity	Investment	23	66	18	66
	Maintenance	11	53	23	18
	Revenue	77	53	77	29
	Price	66	29	29	77
	Demand	8	29	29	43
	Comfort	11	77	23	53
	Historical issues	66	38	43	53
	Security	29	29	66	38
	Infrastructure	53	53	53	43
	Mobility	4	77	43	11
	Legislation	77	43	88	66
Total Green		11	1	6	4
Total Red		5	6	4	8

Source: the authors (2023).

. The number in the cell corresponds to a dispersion ranking position, and these rankings are presented based on color quartiles in order to be more effective. The red cells indicate that the specialists’ evaluation dispersion is larger than the other ties and is in the more discordant quartile. On the other hand, the green cells represent a more consensual evaluation; the white cells represent those in the second and third quartiles. The numbers inside the cells rank the dispersion of pair risk/project. Therefore, evaluating the probability of a lack of comfort risk for P1 was consensual, while for P3, this was wildly divergent when the parties assessed the intensity of the legislation risk.

Table 3 provides some vital information. There is a more consensual evaluation of P1 (total green = 11), probably because P1 represents the palliative project currently used in Maceio. Conversely, project P4 has more divergent results: around 37% of their evaluations are among the 25% more discordant ones. A hypothesis concerning these results is that the project is not clear enough for the specialists, and the pros and cons of the project are diffuse in the general evaluation. Project 4 also involves a problem regarding the takeover of some properties. However, according to the CBTU, this proposal would serve the entire region of Maceió, providing a railroad ring.

This railroad ring will make it possible to expand the coverage area of the UMP, and when the affected region is stabilised, the route can be activated. The regional public transport systems must adopt innovative strategies for demographic changes in the peripheral areas [63], for example, the growing population. The innovative approach can include more sustainable options for transport. Flexible transport systems can reduce car dependency [64].

Table 3 also presents interesting information about the risks involved. The interpretation of the legislation in conformance quality is diffuse and probably requires more discussion concerning the theme. There is a good concordance level of mobility; 62.5% of these risk evaluations are among the 25% more concordant.

As the last step, the Copeland method was applied for all risks, as in Table 1, which computes the number of wins (column W) and losses (row L) and also indicates the Copeland score and Copeland ranking. We performed this procedure for all combinations; thus,

Table 4. Decision results.

General Risk
General			Occurrence			Intensity
Project	Rank	Score	Project	Rank	Score	Project	Rank	Score
P1	3	−1	P1	4	−3	P1	3	−1
P2	4	−3	P2	3	−1	P2	4	−3
P3	1	3	P3	1	3	P3	1	3
P4	2	1	P4	2	1	P4	2	1
Economic Risk
General			Occurrence			Intensity
Project	Rank	Score	Project	Rank	Score	Project	Rank	Score
P1	2	1	P1	1	3	P1	2	1
P2	4	−3	P2	4	−3	P2	4	−3
P3	1	3	P3	2	1	P3	1	3
P4	3	−1	P4	3	−1	P4	3	−1
Social Risk
General			Occurrence			Intensity
Project	Rank	Score	Project	Rank	Score	Project	Rank	Score
P1	4	−3	P1	4	−3	P1	4	−3
P2	3	−1	P2	3	−1	P2	3	−1
P3	1	3	P3	1	3	P3	1	3
P4	2	1	P4	2	1	P4	2	1
Environmental Risk
General			Occurrence			Intensity
Project	Rank	Score	Project	Rank	Score	Project	Rank	Score
P1	4	−3	P1	3	−1	P1	4	−3
P2	3	−1	P2	4	−3	P2	3	−1
P3	1	3	P3	1	3	P3	2	1
P4	2	1	P4	2	1	P4	1	3

Source: The authors (2023).

summarises the rank and scores for each project.

Table 4 helps to clarify the discussion. In this paper, we discuss the project perspective and the issue perspective.

• Social risk is critical in P1. It means that the palliative alternative used in Maceio currently harms transport system users. The economic risk is considerably smaller, probably because the project is in operation, and there are not too many surprises in this evaluation.
• P2 always has the most significant risk, especially economic risks. The specialists do not believe that the project’s financial value corresponds with the project’s actions, and it is probably a project that needs to be avoided.
• P3 had the best ranking in 10 of 12 evaluations. It is comprehensible because P3 represents a combination of the best points between P2 and P4. It is a good reference to compare with P4.
• P4 is a more exciting alternative in terms of risk comparison. This proposal needs improvement in terms of economic risk and must be compared with P3 to improve these results.

The results of these methods were consistent and relevant to the discussion. The technique shows the risks of the project based on a general perspective. It evaluates the sustainability tripod: economic, environmental, and social points of view, assessing the occurrence risk and intensity of the risk.

Another interesting point is that the discordance and relevance computation provided insights into the perspectives of the parties involved. For example, a hypothesis averted in the evaluation is that some parties do not consider the discussion about the time it takes to achieve soil stabilisation, nor was there any discussion about whether there are guarantees that this stability will be achieved. It can generate some divergence, especially in P4 (proposed by the municipal public transport representative). The discussions about this point can diminish bias as a sunk cost, defined by [65] as a situation in which an individual takes into account previously invested resources such as time, money, or effort, i.e., when making a decision, these individuals are influenced by actions (sunk) in the past that will no longer be retrieved and constrain decision making in the present [66,67]. As P3 is precisely the same route as that which was interrupted, we believe that the feeling of being able to recover this investment already made has prevailed. In addition, P3 lacks the sense of returning to normality, abstracting from the fact that there will no longer be the same urban scenario.

4.4. Result Discussion

This paper proposed a method based on some techniques to adequate the model to Maceio’s problem. It is difficult to find similar problems to compare the risk because our study is based on one of the biggest urban tragedies registered. Some articles have similar aspects of this method but with some differences. Ref. [49] presented a problem based on Picture Fuzzy Sets and concepts of social choice based on the Copeland Method to evaluate the risks of a new urban mobility project. The authors did not use the Delphi techniques; the Copeland method was just a general evaluation. Our analysis is profound because the Delphi method enriches the problem structure, and the risk comparison presents more information to identify the conflict process.

Ref. [57] also used the PFS structure but did not use the Delphi scheme and used an evaluation risk in a hotel refrigerating project. They did not mention using any problem structured problem as the Delphi. The authors used a multicriteria method called MABAC to evaluate the PFS numbers. MABAC presents a risk ranking from the most to the least risky. It differentiates our process because our paper indicates, as in Table 3, the conflict degree among the evaluation, and Table 4 shows, based on a sustainable perspective, what are the most relevant points to review. We also calculate these results using the MABAC method [62] just to compare the results with our paper. The rankings obtained by the method were the same. It probably happens because the conditions of this problem are suitable.

The numerical results seem consistent and valuable for this problem. The scores of Table 3 and Table 4 are essentially comparative—a comparison between the analyst and the evaluation criteria. Therefore, the evaluation was not considered absolute but relative to the problems. The case with four projects can bring more ties in evaluation. However, it is not necessarily bad. Also, using a known and precise technique as the Copeland improves the fairness sense about the process. In the cases of a more significant number of specialists or projects, they use Fuzzy aggregators as Ordered Weighted Averaging (OWA) is convenient to handle.

5. Conclusions

Mining activity is generally related to significant socio-environmental impacts as its operations significantly affect the areas in which it is installed. This paper proposed a risk analysis of transport requalification projects in the urban mobility problem caused by a mining disaster related to rock salt extraction in Maceió, Brazil.

The model proposed has three main steps: problem definition, risk management, and decision analysis and uses the Picture Fuzzy-Delphi method for data collection and expert judgment elicitation and the Delphi method to assess the problem without interference from others. In addition, we used Picture Fuzzy Sets (PFSs) to incorporate uncertain information in the decision-making process. The model’s results proved consistent and relevant to the discussion. The results of the application of methods show the risks of the project based on a general perspective. It evaluates the sustainability tripod: economic, environmental, and social points of view, assessing the occurrence risk and intensity of the risk.

The main limitations of this study are related to the methods used to assess risks and the options for requalification projects available at the time of data analysis. Different methods or projects can lead to different results. Another limitation is the sample of stakeholders selected to participate. Future studies can compare the results obtained by applying other methods or still using the same methods with inputs related to other projects.

The post-disaster changes in mobility infrastructure require high investments and the requalification of ongoing mobility options. The difficulty is in implementing these changes without stopping the current model of mobility, because populations need to continue to move in cities. So, the situation generated by the disaster in Maceió is the ongoing chaos in the mobility infrastructure. The local government and the organisations responsible for urban mobility need to carefully analyse the alternatives and the risks involved in its implementation. Despite there being some new and sustainable models of urban mobility, such as bike and car sharing, this is not enough to solve the problem when considering that to enable the new models of mobility it is also necessary to achieve the requalification of the present infrastructure, such as lap bike lanes and roads to bypass congestion. Furthermore, we need to consider that five neighbourhoods with high population occupancy rates had their mobility interrupted, not just reduced. So, the risks presented in this study are useful to apply in similar scenarios and the model proposed connects some traditional techniques of social choice and multicriteria decision aids.

The main contribution of the proposed model is to systematise risks management related to different projects of transport requalification in urban mobility caused by a mining disaster using formal decision-making methods. This finding can be helpful for public decision makers in prioritising projects of mobility and transport requalification with lower risks to the population. Careful decisions need to be made considering the social, environmental, and economic impacts caused by disasters that severely affected the population of Maceió. Therefore, solutions that cause the least impact to the population can be considered by policymakers.